Objective
To compare respiratory compliance and functional residual capacity in infants randomized to a rescue course of antenatal steroids vs placebo.
Study Design
Randomized, double-blinded trial. Pregnant women ≥14 days after initial antenatal steroids were randomized to rescue antenatal steroids or placebo. The primary outcomes were measurements of respiratory compliance and functional residual capacity. This study is registered with clinicaltrials.gov ( NCT00669383 ).
Results
Forty-four mothers (56 infants) received rescue antenatal steroids and 41 mothers (57 infants) received placebo. There was no significant difference in birthweight, or head circumference. Infants in the rescue group had an increased respiratory compliance (1.21 vs 1.01 mL/cm H 2 O/kg; adjusted 95% confidence interval, 0.01-0.49; P = .0433) compared with placebo. 13% in the rescue vs 29% in the placebo group required ≥30% oxygen ( P < .05). Patients delivered at ≤34 weeks had greater pulmonary benefits.
Conclusion
Infants randomized to rescue antenatal steroids have a significantly increased respiratory compliance compared with placebo.
Respiratory distress syndrome (RDS) is a common complication of preterm delivery and increases neonatal mortality and morbidity. A single course of antenatal steroids (AS) remains the standard of care for women at risk for preterm delivery between 24 and 34 weeks of gestation, and significantly decreases mortality and RDS. The optimal response to a course of AS occurs if it is given at least 24 hours before but within 7 days of delivery. The duration of the clinical effect of a course of AS remains uncertain, and the efficacy and safety of repeat courses of AS remains a topic of continuing debate and study.
For Editors’ Commentary, see Table of Contents
Recent randomized data show weekly courses of AS in preterm infants may improve lung function outcomes. Crowther et al randomized 982 women at <32 weeks of gestation to weekly AS vs placebo. They demonstrated a significant decrease in RDS and severe lung disease in the AS group. The Eunice Kennedy Shriver National Institute of Child Health and Human Development Maternal Fetal Medicine Units Network also reported improved pulmonary outcomes in preterm infants randomized to weekly courses of AS vs placebo. Despite the suggestion of pulmonary benefits after repeat courses of AS, concern remains of possible harm after weekly AS therapy, particularly in terms of reduced in utero growth and long-term neurodevelopmental outcome. However, a recent trial by Garite et al of 1 rescue course of AS vs placebo showed improved neonatal outcome without apparent increased short-term risk.
We have used measurements of pulmonary function, specifically passive respiratory compliance (Crs) and functional residual capacity (FRC), as an objective and reproducible way of quantifying the effects of AS on pulmonary function. These physiologic measurements correlate with clinical respiratory outcomes. We recently reported that infants ≤32 weeks of gestation treated with a single course of AS >7 days before delivery had a significantly lower Crs compared with matched infants treated with a course of AS 1-7 days before delivery. The difference in Crs was more striking if the infants received AS >14 days before delivery. Our objective was to compare pulmonary function (Crs and FRC) in infants who received a course of AS, remained undelivered for at least 14 days, and were randomized to a single rescue course of AS vs placebo. We hypothesized that infants who received a rescue course of AS would have a significantly increased Crs compared with those who received placebo.
Materials and Methods
Eligibility criteria
This prospective randomized trial was conducted at the Neonatal Intensive Care Unit at Oregon Health and Science University (Portland, OR) and at Sacred Heart Hospital (Pensacola, FL). The study was approved by the institutional review boards at each institution and informed consent obtained for each enrolled patient. This study is registered with clinicaltrials.gov (registration no. NCT00669383 ). Inclusion criteria included: (1) pregnant women 26 to <34 weeks of gestation; (2) at least 14 days after first course of AS (93% received betamethasone); (3) continued risk of preterm delivery as determined by their care provider; and (4) informed consent.
We excluded patients with: (1) multiple gestations greater than twins; (2) insulin dependent diabetes; (3) clinical chorioamnionitis; (4) major documented fetal or chromosomal abnormalities; (5) first course of AS given at <24 weeks of gestation; and (6) chronic steroid use during pregnancy for clinical care.
Study design
This was a prospective, randomized, double-blinded trial stratified by gestational age (≤28 vs >28 weeks’ gestation) and multiple gestation (twins vs singletons). After obtaining informed consent, patients were randomized to rescue AS or placebo only if there were recurrent signs of preterm delivery as determined by the mother’s clinician. Patients assigned to the “rescue AS” group received a course of AS (two 12 mg intramuscular injections of betamethasone [Celestone Soluspan; Schering-Plough Corporation, Kenilworth, NJ] 24 hours apart). Those assigned to the “placebo” group received 2 doses of placebo that were identical in appearance to the betamethasone and consisted of 25 mg cortisone acetate, an inactive steroid. Group assignment was performed using a randomization table. A staff pharmacist performed randomization and study drug preparation at each institution, and all patients, investigators, and care providers were unaware of treatment allocation. Pulmonary function was measured within 72 hours of birth and before surfactant therapy, when required. Infants were studied in the supine position while quietly asleep; no sedation was used.
Comparisons of Crs and FRC between groups were our primary endpoints. Other pertinent clinical outcome measures between groups were also monitored. All outcomes were assessed blinded to treatment group allocation.
Measurements
Respiratory mechanics and lung volume were measured with a computerized infant pulmonary function cart (SensorMedics 2600; SensorMedics Inc, Yorba Linda, CA). Crs was obtained with the single-breath occlusion technique and FRC with the nitrogen washout method. In intubated patients, including those requiring surfactant, testing was performed by connecting the infant’s endotracheal tube into the system via a 3-way valve that also connected to the ventilator. In nonintubated patients, a face mask was used.
For the single-breath occlusion technique, the airway was briefly occluded at end inspiration until an airway pressure plateau was observed and the Hering Breuer reflex invoked. The linear portion of the passive flow-volume curve was identified, and a regression line drawn to obtain the best fit. From the intercepts on the flow and volume axes, respiratory system compliance and resistance were calculated. Acceptance criteria included: (1) stable end expiratory baseline; (2) plateau pressure lasting >100 msec; (3) plateau pressure varying by < ± 0.125 cm H 2 O; (4) acceptable flow-volume curve by visual inspection, with linear data segment identified; and (5) at least 10 breaths accepted with a coefficient of variation of <20%.
For the nitrogen washout technique, calibration was performed with 2 known volumes, and a calibration line was constructed for the system at the specific flow rate. The calibration curve was then used to correlate the nitrogen washed out to the infant’s FRC. The system corrected for dead space present and corrected the FRC to body temperature, pressure, and water-saturated conditions. Total FRC was related to body weight. Acceptance criteria included: (1) infant supine and quietly asleep; (2) test initiated at end expiration; (3) no evidence of leak on tracing of the washout; (4) consistent tracings; and (5) at least 3 measurements with a coefficient of variation of <10%.
Growth measurements (weight, head circumference, and length) at birth and hospital discharge were compared between groups, and corresponding Z-scores for these measurements were calculated. Other clinical outcome parameters, including surfactant administration, diagnosis of RDS (defined as clinical signs of respiratory distress with radiographic appearance and needing supplemental oxygen with FiO 2 >0.21), respiratory distress with FiO 2 requirement ≥0.30 and ≥0.40 at 24 hours of age, days on mechanical ventilation, and days on supplemental oxygen were monitored.
Statistical methods
Differences in continuous variables between the 2 groups were first analyzed by 2-tailed, independent samples Student t tests or the Mann-Whitney U test where appropriate (for data not normally distributed). Categorical variables were evaluated with χ tests and Fisher’s exact test where appropriate. Data are presented as mean ± standard deviation (SD), unless otherwise indicated.
The statistical analyses were on intention-to-treat. Comparisons of the primary outcomes between groups were performed using linear mixed modeling. This approach accounts for the correlation (nonindependence) between twins, and allows for adjustment of additional covariates and potential confounders. We began with a model, including rescue dose, gestational age (GA) at birth, multiple gestations, maternal smoking history, rupture of membranes, and gestational diabetes, as these were the initial predictors of interest. The remaining covariates: GA at first AS dosing, GA at study dosing, race, gender, and mode of delivery were introduced to the model 1 at a time, and examined for their statistical significance and/or their effect on the relationship between rescue dose and each outcome. We also planned a subgroup analysis for infants who delivered at ≤34 weeks of gestation. We used SPSS for Windows, Version 15 (SPSS, Inc, Chicago, IL) and SAS 9.1.3 (SAS Institute Inc, Cary, NC) for analyses.
Power calculations were based on previous studies. We reported a 50% increase in Crs in preterm infants treated with a single course of AS 1-7 days before delivery when compared with matched untreated infants. We also conducted a randomized trial, in which infants undelivered 1 week after their initial AS course were randomized to weekly AS versus placebo. Infants randomized to placebo delivered an average of 24 days after AS dosing and had Crs measurements that were on average 20% lower than in infants randomized to weekly AS. These findings were not significantly different perhaps due to small sample size, as the study was stopped early due to concerns of possible adverse effects of weekly AS therapy.
For our current study, we hypothesized that infants in the rescue AS group would have significantly increased Crs than those in the placebo group. We estimated that to show about a 30% difference in Crs between the groups, we would need to study approximately 40 pregnancies in each group to reject the null hypothesis with a type I error of 0.05 and a power of 80%. Data were reviewed once by the independent, data safety monitoring committee, who were not aware of treatment-group allocation. An interim analysis was conducted when 49.6% (56/113) of the infants were delivered.
Results
Patients were recruited from June 2001 to May 2007. The study profile is shown in the Figure . A total of 44 mothers with 56 infants were randomly assigned to a rescue course of AS and 41 mothers with 57 infants to placebo. The first 8 patients were recruited at Sacred Heart Hospital, with the remaining patients recruited at Oregon Health and Science University. Eighty-six percent of the rescue and 90% of the placebo group received a full course of the study medication. Neonatal data were not available for 1 infant in the placebo group, resulting in 56 vs 56 infants available for analysis ( Figure ).
Maternal and infant characteristics were comparable between the groups ( Tables 1 and 2 ), except for significantly more smokers by history in the rescue group. Both groups of mothers received their first course of AS at about 27 weeks; and received their study dose at about 30 weeks of gestation ( Table 1 ). Patients in the rescue AS group delivered 8 (2-24) days after receiving rescue/study dosing vs 11 (3-20) days in the placebo group (median, 25th-75th percentiles; P = .66). Eighty-three of a total of 113 infants (73.5%) were delivered at ≤34 weeks. There were no significant differences in birthweight, GA at birth, admission head circumference, birth length, or incidence of small for GA infants (birthweight <10th percentile) between groups. There was no difference in Z-scores for birthweight, head circumference, or length between groups ( Table 2 ).
| Characteristic | Rescue AS (n = 44) | Placebo (n = 41) | P |
|---|---|---|---|
| Maternal age, y a | 26.9 ± 7.5 | 28.6 ± 6.4 | NS |
| White, n (%) | 30 (68) | 27 (66) | NS |
| Maternal smoking, n (%) | 11 (25) | 2 (5) | < .05 |
| Twin gestation, n (%) | 12 (27) | 16 (39) | NS |
| Rupture of membranes, h b | 0.5 (0-9) | 0 (0-8) | NS |
| Rupture of membranes >24 h, n (%) | 9 (20) | 6 (15) | NS |
| Gestational diabetes, n (%) | 2 (5) | 5 (12) | NS |
| Cesarean section, n (%) | 24 (55) | 24 (59) | NS |
| GA at 1st AS dosing, wk a | 26.6 ± 1.9 | 27.1 ± 2.2 | NS |
| GA at study dosing, wk a | 29.8 ± 1.9 | 30.3 ± 2.1 | NS |
| Preterm labor, n (%) | 35 (80) | 29 (71) | NS |
| Preeclampsia, n (%) | 3 (7) | 2 (5) | NS |
| Antepartum hemorrhage, n (%) | 7 (16) | 11 (27) | NS |
| Characteristic | Rescue AS (n = 56) | Placebo (n = 56) | P |
|---|---|---|---|
| GA at birth, wk a | 31.9 ± 3.3 | 32.3 ± 2.9 | NS |
| White, n (%) | 39 (70) | 38 (68) | NS |
| Female, n (%) | 28 (50) | 29 (52) | NS |
| Apgar score 1, 5 min b | 7, 8 | 8, 9 | NS |
| Birthweight, g a | 1806 ± 778 | 1830 ± 657 | NS |
| Birthweight Z score a | −0.14 ± 0.86 | −0.14 ± 0.98 | NS |
| Small for gestational age, n (%) | 6 (11) | 8 (14) | NS |
| Birth head circumference, cm a,c | 28.7 ± 3.0 | 29.2 ± 2.9 | NS |
| Birth head circumference Z score a | −0.19 ± 0.90 | −0.18 ± 0.93 | NS |
| Birth length, cm a | 41.9 ± 4.7 | 42.4 ± 4.4 | NS |
| Birth length Z score a | −0.11 ± 0.99 | −0.06 ± 0.96 | NS |
Pulmonary function tests were performed at a median postnatal age of 22 hours in the rescue AS and 21 hours in the placebo group. Successful measurements of Crs were obtained in 49 rescue and 49 placebo patients. Accounting for the correlation between twins and using linear mixed modeling, there was a significant difference in Crs with the rescue AS group having a mean Crs of 1.21 mL/cm H 2 O/kg vs 1.01 mL/cm H 2 O/kg in the placebo group, adjusted 95% confidence interval (CI) for difference (0.01-0.49), P = .0433 ( Table 3 ). We did not demonstrate a significant difference in FRC between groups: 24.8 mL/kg in the rescue vs 22.0 mL/kg in the placebo group, adjusted 95% CI for difference (−1.40 to 6.62), P = .19 ( Table 3 ).
